Abstract
Understanding how potential climate change will affect availability of water resources for citrus production globally is needed. The main goal of this study is to investigate impacts of potential future climate change on citrus irrigation requirements (IRR) in major global citrus producing regions, e.g., Africa, Asia, Australia, Mediterranean, Americas. The Irrigation Management System (IManSys) model was used to calculate optimum IRR for the baseline period (1986–2005) and two future periods (2055s and 2090s) subject to combination of five and seven temperature and precipitation levels, respectively. Predicted IRR show significant spatio-temporal variations across study regions. Future annual IRR are predicted to globally decrease; however, future monthly IRR showed mixed results. Future evapotranspiration and IRR are projected to decrease by up to 12 and 37%, respectively, in response to increases in CO2 concentration. Future citrus canopy interception and drainage below citrus rootzones are expected to slightly increase. Annual rainfall changes are negatively correlated with changes in IRR. These projections should help the citrus industry better understand potential climate change impacts on citrus IRR and major components of the water budget. Further studies are needed to investigate how these potential changes in CO2 concentration, temperature, evapotranspiration, rainfall, and IRR will affect citrus yield and its economic impact on the citrus industry.
INTRODUCTION
Citrus is one of the major fruit crops covering significant agricultural areas globally (Liu et al. 2012). While global citrus production was projected to decrease in 2016 compared with 2015 (USDA 2016), in 2015, Brazil, China, and the USA were the top citrus fruit producing countries in the world (USDA 2015). The citrus industry contributes considerably to national gross domestic products (GDPs) of several countries. For example, in the state of Florida, Hodges et al. (2014), in their economic impact analysis study, estimated that citrus fruit production contributed 3.82 billion USD from a total of 156 million boxes of citrus fruit produced during the 2012/2013 production season. While the citrus industry signficantly contributes to local, national, and global economies, its production consumes considerable amounts of fresh water resources, which makes the industry compete for available fresh water resources with other major water users, e.g., production of other crops, domestic and industrial uses, and ecosystems services.
In addition, the citrus industry is also facing critical challenges, including the wide spread of diseases and pests (e.g., citrus greening, citrus canker), which results in major damage to citrus production and in economic losses (Schubert et al. 2001; Gottwald et al. 2002; Das 2003; Hodges & Spreen 2006). For example, the 2015/16 global orange production was predicted to decrease by 3.0 million metric tons, where significant yield reduction was expected in some of the major producing regions (e.g., Brazil, USA, and South Africa) (USDA 2016). Moreover, the interaction of projected increases in temperature and reduction in rainfall, in most parts of the globe, due to climate change will certainly have a significant impact on available fresh water resources. Expected increases in climate variability, severe droughts, and recurrent floods will certainly have both direct and indirect effects on the environment and natural resources (Hulme et al. 1999; McGeehin & Mirabelli 2001; Haines et al. 2006; O'Sullivan 2015). The hydrologic cycle will be likely among the key ecosystem processes that will be adversely affected by climate change (Arnell 1999; Agarwal et al. 2014; Hassan et al. 2014). This will lead to more competition for available fresh water resources between different water user groups, which could be even further aggravated by uncertainties in the extent and intensity of climate change.
Climate change and variability not only affect available water resources directly but also affect crop physiological characteristics, mainly due to increases in CO2 concentration, which in turn affect photosynthetic processes. Understanding the potential impacts of increased CO2 concentration in the atmosphere on plant physiology and thus agriculture has attracted attention (Allen et al. 1991; Wullschleger et al. 2002; Ficklin et al. 2010; Hussain et al. 2013), and studies have shown that an increase in atmospheric CO2 concentration not only results in an increase in air temperature but also affects leaf stomata conductance and consequently evapotranspiration processes (Allen et al. 1998, 1991; Ficklin et al. 2010; Hussain et al. 2013; Fares et al. 2016).
In order for the citrus industry to continue to be economically viable, there is a need to better understand how potential future climate would affect availability of fresh water resources for the industry in the mid and long future periods: 2055s (2046–2065) and 2090s (2081–2100) compared with the baseline period (1986–2005) (Du Plessis 1985; Morgan et al. 2007; Villalobos et al. 2009). Such information would shed more light on the extent of potential challenges and help the citrus industry plan short-term mitigation measures and long-term adaptation strategies.
Understanding how potential climate change affects the availability of water resources and major components of the water budget is critical (Ficklin et al. 2010; Fares et al. 2016). Therefore, the objectives of this study are to (1) calculate current citrus irrigation requirements and (2) predict how potential climate change scenarios would affect citrus IRR and other field water budget components, e.g., effective rainfall (ER), evapotranspiration (ETo), canopy interception (INT), drainage (DR), and runoff (RO) across the major producing regions in the world.
MATERIALS AND METHODS
Study sites
This study was conducted in selected locations of the major citrus producing regions across the world: Africa (Cape Town, South Africa), Asia (Mersin, Turkey), Australia (Riverland, Australia), Mediterranean (Nabeul, Tunisia), North America (Riverside, California; Fort Pierce and Lake Alfred, Florida; and Brownsville, Texas), and South America (Sao Paulo, Brazil) (Figure 1). For this work, the study sites were chosen with the underlying objective to represent most of the major citrus producing regions according to the report by the United States Department of Agriculture (USDA 2016). However, it is also worth noting that this study does not include some of the leading citrus producing countries (e.g., China).
Map of the locations used in this study which represent the major citrus producing areas across the world.
Map of the locations used in this study which represent the major citrus producing areas across the world.
Input data
Daily precipitation, minimum temperature (Tmin), maximum temperature (Tmax), and wind speed data were obtained from the National Centers for Environmental Prediction – Climate Forecast System Reanalysis (NCEP–CFSR) global weather data (http://globalweather.tamu.edu/) for the baseline period (1986–2005). The baseline period was chosen based on the Intergovernmental Panel on Climate Change – Fifth Assessment Report (Stocker et al. 2013). Daily evapotranspiration data were computed using the modified FAO Penman–Monteith equation (Allen et al. 1998), built in the IManSys model (discussed in detail in the section below), which accounts for changes in CO2 concentration.

Available water content was calculated as the difference between water content at field capacity (FC: θp at 33 kPa) and permanent wilting point (PWP: θp at 1,500 kPa).
Soil information used to calculate soil hydraulic properties was obtained from the FAO harmonized world soil database v 1.2 (http://www.fao.org/soils-portal/soil-survey/soil-maps-and-databases/harmonized-world-soil-database-v12/en/).
Irrigation Management System (IManSys) model






IManSys model inputs of soil, crop, and irrigation parameters
Location . | Parameter . | |
---|---|---|
Irrigation system . | Soil type . | |
Brownsville | Flood | Laredo |
Lake Alfred | Micro sprinkler | Candler sand |
Fort Pierce | Micro sprinkler | Mayakka |
Nabeul | Drip | Calcaric Regosols |
Sao Paulo | Drip | Haplic Ferralsols |
Cape Town | Drip | Lithic Leptosols |
Riverland | Drip | Calcic Xerosols |
Riverside | Drip | Greenfield |
Mersin | Drip | Chromic Luvisols |
Location . | Parameter . | |
---|---|---|
Irrigation system . | Soil type . | |
Brownsville | Flood | Laredo |
Lake Alfred | Micro sprinkler | Candler sand |
Fort Pierce | Micro sprinkler | Mayakka |
Nabeul | Drip | Calcaric Regosols |
Sao Paulo | Drip | Haplic Ferralsols |
Cape Town | Drip | Lithic Leptosols |
Riverland | Drip | Calcic Xerosols |
Riverside | Drip | Greenfield |
Mersin | Drip | Chromic Luvisols |



Climate change scenarios
Potential future climate change scenarios for two future periods 2055s (2046–2065) and 2090s (2081–2100) were generated based on the baseline climate records that cover the period 1986–2005. Six precipitation levels were prepared for the two future periods by adjusting daily precipitation records of the baseline period by ±5, ±10, and ±20%. Similarly, baseline minimum and maximum temperature records were adjusted by adding +1, +1.4, and +1.3 °C in 2055s and +1, +1.8, and +2.2 °C in 2090s to reflect projected changes in global mean temperature under three representative concentration pathways (RCPs): RCP 2.6, RCP 4.5, and RCP 6.0 of the Intergovernmental Panel on Climate Change – Fifth Assessment Report (IPCC–AR5), respectively. The RCPs represent changes in radiative forcing values of +2.6, +4.5, and +6.0 W m−2, respectively (Table 2).
Projected increases in mean air temperature under different RCPs and corresponding CO2 levels for two future periods
RCP . | Change in mean temperature (°C) . | CO2 equiv. (ppm) . | Key features . | |
---|---|---|---|---|
2055s . | 2090s . | |||
RCP 2.6 | 1.0 | 1.0 | 490 | Peak and decline, which leads to very low greenhouse gas concentration |
RCP 4.5 | 1.4 | 1.8 | 650 | Stabilization scenario, radiative forcing is stabilized before 2100 |
RCP 6.0 | 1.3 | 2.2 | 850 | Stabilization scenario, radiative forcing is stabilized after 2100 |
RCP . | Change in mean temperature (°C) . | CO2 equiv. (ppm) . | Key features . | |
---|---|---|---|---|
2055s . | 2090s . | |||
RCP 2.6 | 1.0 | 1.0 | 490 | Peak and decline, which leads to very low greenhouse gas concentration |
RCP 4.5 | 1.4 | 1.8 | 650 | Stabilization scenario, radiative forcing is stabilized before 2100 |
RCP 6.0 | 1.3 | 2.2 | 850 | Stabilization scenario, radiative forcing is stabilized after 2100 |
Sources:Moss et al. (2010) and Stocker et al. (2013).
RESULTS AND DISCUSSION
Climate characteristics
Analysis of 20 years (1986–2005) minimum and maximum temperatures, and precipitation provided useful information about climate variables in the study locations. As expected, the selected study sites have different climates (Figure 2). Annual rainfall distribution was not uniform across the months; most of the locations have a unimodal rainfall distribution (Figure 2). Sao Paulo receives the highest rainfall (2,016 mm yr−1) (Table 3) and monthly rainfall was greater than corresponding evapotranspiration (Figure 2). Our results for Sao Paulo concur with the reports of other studies (e.g., Liebmann et al. 2001; Silva Dias et al. 2013; Coelho et al. 2016). Riverside receives the smallest rainfall (332 mm yr−1) and ETo was significantly greater than rainfall for about ten months of the year (Figure 2). Annual ETo was greatest in Brownsville with 1,641 mm yr−1, while the smallest ET was in Cape Town (1,080 mm yr−1). Overall, annual rainfall was smaller than annual evapotranspiration for six of the nine locations of the study, which in turn would lead to a deficit in soil moisture unless supplied with irrigation (Table 3). This is in agreement with a previous work that reported ETo as the major component of the hydrologic cycle (Liu et al. 2010). On the other hand, recent studies also show that irregular precipitation distribution throughout the growing season results in water deficit that could be a critical issue even in regions where rainfall is comparatively higher such as Sao Paulo (Coelho et al. 2016).
Annual average summary of climate characteristics of studied locations (based on 1986–2005 data)
Location . | Tmin (°C) . | Tmax (°C) . | P (mm) . | ETo (mm) . | Deficit (P-ETo) (mm) . |
---|---|---|---|---|---|
Brownsville | 20.7 | 29.5 | 1,023.4 | 1,641.1 | −617.7 |
Cape Town | 13.8 | 20.7 | 651.3 | 1,079.8 | −428.5 |
Fort Pierce | 22.2 | 29.0 | 1,931.9 | 1,423.4 | 508.5 |
Lake Alfred | 17.2 | 27.7 | 1,972.3 | 1,412.0 | 560.3 |
Mersin | 8.4 | 18.9 | 973.6 | 1,166.4 | −192.8 |
Nabeul | 15.9 | 22.7 | 531.4 | 1,165.9 | −634.5 |
Riverland | 11.4 | 24.0 | 367.3 | 1,344.9 | −977.6 |
Riverside | 10.0 | 25.1 | 331.7 | 1,452.4 | −1,120.6 |
Sao Paulo | 15.0 | 23.9 | 2,015.5 | 1,288.5 | 726.9 |
Location . | Tmin (°C) . | Tmax (°C) . | P (mm) . | ETo (mm) . | Deficit (P-ETo) (mm) . |
---|---|---|---|---|---|
Brownsville | 20.7 | 29.5 | 1,023.4 | 1,641.1 | −617.7 |
Cape Town | 13.8 | 20.7 | 651.3 | 1,079.8 | −428.5 |
Fort Pierce | 22.2 | 29.0 | 1,931.9 | 1,423.4 | 508.5 |
Lake Alfred | 17.2 | 27.7 | 1,972.3 | 1,412.0 | 560.3 |
Mersin | 8.4 | 18.9 | 973.6 | 1,166.4 | −192.8 |
Nabeul | 15.9 | 22.7 | 531.4 | 1,165.9 | −634.5 |
Riverland | 11.4 | 24.0 | 367.3 | 1,344.9 | −977.6 |
Riverside | 10.0 | 25.1 | 331.7 | 1,452.4 | −1,120.6 |
Sao Paulo | 15.0 | 23.9 | 2,015.5 | 1,288.5 | 726.9 |
Seasonal climate characteristics of study sites for the baseline period (1986–2005).
Seasonal climate characteristics of study sites for the baseline period (1986–2005).
Model outputs
Evapotranspiration is the major component of the water budget in all study sites (Figure 3). Moreover, owing to variations in climatic parameters (rainfall, evapotranspiration, and temperature), effective rainfall (ER), gross irrigation requirement (IRR), and drainage (DR) showed significant differences across the studied locations. Evapotranspiration results are in agreement with findings of other studies that reported evapotranspiration as the major component of the hydrologic cycle (Liu et al. 2010). Overall, effective rainfall, gross irrigation requirement, and evapotranspiration are predicted to decrease in the 2055s and 2090s compared with the baseline period, regardless of geographic location. However, while the reduction in ER, IRR, and ETo were significant in 2055s and 2090s compared to baseline, the differences between the two future periods (2055s and 2090s) were relatively small (Figure 3). On the other hand, canopy interception and drainage are expected to show a slight increase during the two future periods compared with the baseline. Runoff losses were insignificant across locations. Overall, Riverside has the greatest IRR, while Sao Paulo had the smallest during the baseline period; this trend, for IRR, is projected to continue in the mid and end of the 21st century. Our results were in agreement with Fares et al. (2016), who reported a reduction in ETo and IRR with an increase in CO2 concentrations for seed corn and coffee in Hawaii. Similarly, several studies reported a projected reduction in ETo with increases in CO2 concentration (Allen et al. 1991; Lockwood 1999). Hussain et al. (2013) also reported that elevated CO2 concentration resulted in a significant reduction in ETo for a C4 crop (maize) compared to a C3 crop (soybean). This finding concurs with previous reports by several authors who argued that an increase in atmospheric CO2 concentration would trigger plant stomata closure and thereby lead to a reduction in ETo (Medlyn et al. 2001; Wullschleger et al. 2002; Shams et al. 2012). Medlyn et al. (2001), in their meta-analysis study, showed that stomatal conductance of woody species showed up to a 21% reduction in response elevated CO2 concentration. Allen et al. (2011) found that elevated CO2 concentration enhanced water use efficiency of maize and sorghum plants at early crop growth season. They suggested that CO2 concentration could potentially ameliorate drought stress on C4 crops. Similarly, Kimball (2016) reported that an incrase in CO2 concentration from 353 to 550 ppm resulted in a 10% reduction in evapotranspiration. In contrast, however, studies in different parts of the world projected increases in ETo and IRR in response to CO2 concentration increase. Fader et al. (2016) estimated that gross irrigation requirements would increase between 4 and 18% due to climate change in the Mediterranean regions by 2085s. Similarly, Lee & Huang (2014) and Rodríguez Díaz et al. (2007) reported increases in irrigation requirement, due to climate change, in northern Taiwan and Spain, respectively.
Model output summary for major water budget components by study site. Results are based on combined temperature and precipitation scenarios.
Model output summary for major water budget components by study site. Results are based on combined temperature and precipitation scenarios.
Moreover, monthly IRR reduction will not occur at a constant rate throughout the year and across the study sites (Figure 4(a)). While overall average IRR is projected to decrease in the 2055s and 2090s compared with the baseline, there will also be some increases in monthly IRR during some months of the year in most study sites (e.g., during November in Sao Paulo, and February and November in Nabeul) (Figure 4(a)). In Nabeul, Tunisia, IRR is predicted to increase by up to 100% during February and November both in 2055s and 2090s, while during January IRR is predicted to decrease by up to 50% (Figure 4(a)). Similarly, a greater decrease in ETo is predicted in the 2055s than in the 2090s compared with the baseline period (Figure 4(b)). Additionally, while ETo is projected to consistently decrease throughout the year in all study sites, the magnitude of reduction varies between months (Figure 4(b)). Overall, greater reductions in IRR and ETo are projected in the 2055s compared to 2090s with respect to the baseline.
Predicted changes in monthly irrigation requirement (IRR) (a) and evapotranspiration (ETo) (b) for the two future periods (2055s and 2090s) compared with the baseline (1986–2005).
Predicted changes in monthly irrigation requirement (IRR) (a) and evapotranspiration (ETo) (b) for the two future periods (2055s and 2090s) compared with the baseline (1986–2005).
Seasonal variations of irrigation requirement and evapotranspiration between representative concentration pathways
Projected changes in IRR and ETo between RCPs show great seasonal variability (Figure 5(a) and 5(b)).
Differences in monthly changes in irrigation requirement (IRR) (a) and evapotranspiration (ETo) (b) between representative concentration pathways (RCPs).
Differences in monthly changes in irrigation requirement (IRR) (a) and evapotranspiration (ETo) (b) between representative concentration pathways (RCPs).
Significant decreases in both IRR and ETo are predicted to occur at higher CO2 concentrations (RCP 6.0) compared with the baseline, wherein most locations' IRR and ETo are predicted to have smaller values than the baseline, except a few exceptions for IRR where slight increases in IRR were projected during some months of the year (during November in Mersin, and February and November in Nabeul) compared with the baseline. Our results are in agreement with the findings of Allen et al. (1991) and Fares et al. (2016), who observed a significant reduction in ETo and IRR under elevated CO2 concentrations. However, on the other hand, Wullschleger et al. (2002) argued that hydrological response of plants to CO2 concentration is affected by the temporal and spatial scale of observation.
Overall, while future annual average IRR and ETo are projected to show slight decreases under RCP 2.6, their values are within the same order of magnitude as those for the baseline, especially in the humid subtropical areas of the globe, i.e., Fort Pierce, Lake Alfred, and Sao Paulo. The significant reduction of ETo in response to elevated CO2 concentration is clearly shown in Figure 6, where significantly smaller ETo was predicted under the highest CO2 concentration scenario (RCP 6.0, CO2 = 850 ppm) of this study. Ramirez & Finnerty (1996) argued that elevated CO2 concentration will have a beneficial effect on irrigated agriculture by improving water use efficiency of crops.
Correlation between baseline evapotranspiration and predicted evapotranspiration under different representative concentration pathways (RCPs). Results are based on combined data from all precipitation scenarios.
Correlation between baseline evapotranspiration and predicted evapotranspiration under different representative concentration pathways (RCPs). Results are based on combined data from all precipitation scenarios.
Effect of temperature on major components of the water budget
Under the baseline greenhouse gas emission scenario (with an average CO2 concentration of 360 ppm), an increase in temperature will result in an increase in ETo and thereby IRR, regardless of the geographic location of the study (Table 4). In agreement with this, Allen et al. (1998) reported that air temperature, among other climatic variables, regulates ETo by controlling moisture-holding capacity of the air and soil vapor fluxes. Similarly, Allen (1999) reported that evapotranspiration will increase about 4 to 5% per 1 °C rise in temperature under well-watered conditions. However, the effect of an increase in temperature was greater in humid areas (e.g., Sao Paulo) where an increase in average temperature by 1 °C will result on average in an 8% increase of IRR. The effect of temperature was smaller in Riverside and Riverland where an average increase of IRR by 3% is projected due to an increase in average temperature by 1 °C. A study by Vara Prasad et al. (2005) showed that while elevated CO2 concentration increases crop yield under normal conditions, the beneficial effects of elevated CO2 are signficantly offset by negative effects of high temperature on yield and yield-components beyond optimum temperature.
Effect of temperature on citrus irrigation requirement under current (baseline) greenhouse emission scenario (360 ppm)
Location . | Baseline IRR (mm) . | Change in annual IRR (%) . | ||||
---|---|---|---|---|---|---|
+ 1 °C . | + 1.3 °C . | + 1.4 °C . | + 1.8 °C . | + 2.2 °C . | ||
Brownsville | 1,208 | 4.1 | 5.3 | 5.8 | 7.6 | 9.5 |
Cape Town | 874 | 4.5 | 5.8 | 6.8 | 8.1 | 9.6 |
Fort Pierce | 522 | 6.1 | 8.0 | 8.4 | 10.9 | 12.5 |
Lake Alfred | 573 | 5.1 | 6.5 | 7.2 | 9.4 | 11.5 |
Mersin | 804 | 3.1 | 5.2 | 6.6 | 8.5 | 9.5 |
Nabeul | 999 | 3.4 | 5.4 | 6.2 | 7.4 | 8.7 |
Riverland | 1,315 | 3.3 | 4.2 | 4.4 | 5.9 | 7.5 |
Riverside | 1,535 | 2.9 | 3.7 | 4.2 | 5.3 | 6.5 |
Sao Paulo | 353 | 9.1 | 10.8 | 10.8 | 13.6 | 13.3 |
Location . | Baseline IRR (mm) . | Change in annual IRR (%) . | ||||
---|---|---|---|---|---|---|
+ 1 °C . | + 1.3 °C . | + 1.4 °C . | + 1.8 °C . | + 2.2 °C . | ||
Brownsville | 1,208 | 4.1 | 5.3 | 5.8 | 7.6 | 9.5 |
Cape Town | 874 | 4.5 | 5.8 | 6.8 | 8.1 | 9.6 |
Fort Pierce | 522 | 6.1 | 8.0 | 8.4 | 10.9 | 12.5 |
Lake Alfred | 573 | 5.1 | 6.5 | 7.2 | 9.4 | 11.5 |
Mersin | 804 | 3.1 | 5.2 | 6.6 | 8.5 | 9.5 |
Nabeul | 999 | 3.4 | 5.4 | 6.2 | 7.4 | 8.7 |
Riverland | 1,315 | 3.3 | 4.2 | 4.4 | 5.9 | 7.5 |
Riverside | 1,535 | 2.9 | 3.7 | 4.2 | 5.3 | 6.5 |
Sao Paulo | 353 | 9.1 | 10.8 | 10.8 | 13.6 | 13.3 |
Results are based on combined data from all precipitation change scenarios for the baseline period.
It is worth mentioning that in humid areas where overall IRR is smaller (e.g., Sao Paulo), a small increase in IRR will result in greater percentage changes (Table 4 and Figure 7(a)). Similarly, an increase in temperature under the baseline CO2 concentration (360 ppm) will result in an increase in ETo (Figure 7(b)), while DR and INT will slightly decrease in some places or remain constant (Figure 7(c) and 7(d)). Our findings concur with those of Fares et al. (2016), who found similar results for different crops (seed corn and coffee).
Effect of temperature on projected major water budget components in the 2055s and 2090s: (a) irrigation requirement (IRR), (b) evapotranspiration (ETo), (c) drainage (DR), and (d) canopy interception (INT). Results are based on combined data from all precipitation scenarios.
Effect of temperature on projected major water budget components in the 2055s and 2090s: (a) irrigation requirement (IRR), (b) evapotranspiration (ETo), (c) drainage (DR), and (d) canopy interception (INT). Results are based on combined data from all precipitation scenarios.
Nevertheless, in the two future periods (2055s and 2090s), an increase in CO2 concentration will mask the effect of temperature and will result in an overall decrease in IRR and ETo (Figure 7(a) and 7(b)), while DR and INT will increase (Figure 7(c) and 7(d)).
However, the magnitude of increase in DR and INT due to an increase in CO2 concentration in the 2055s and 2090s varies significantly between study sites, where greater increases are projected in the humid areas of the study, Fort Pierce, Lake Alfred, and Sao Paulo compared with the other locations (Figure 7(c) and 7(d)). This clearly shows that the magnitude of the effect of climate change varies depending on the climate condition of the location under consideration.
Effect of rainfall on major components of the water budget
In general, as expected, reduction in rainfall will result in an increase in IRR across all study sites (Figure 8(a)). However, a change in rainfall will have a significant effect on IRR in humid regions (i.e., Fort Pierce, Lake Alfred, and Sao Paulo), where, for example, in Sao Paulo, a 20% increase in rainfall would result in a 40% reduction in IRR under RCP 6.0. In contrast, however, in arid areas (e.g., Riverside), the effect of rainfall is relatively negligible compared with other locations. Surprisingly, a 20% reduction in rainfall under higher CO2 concentration (RCP 6.0) scenario will still result in a decrease in IRR compared with that of the baseline. This indicates, at higher CO2 concentrations, the effect of CO2 is dominant over the change in rainfall. This is also clearly visible in Figure 8(b), where the effect of rainfall was masked by higher CO2 concentrations, and changes in ETo were only observed between RCPs. On the other hand, an increase in rainfall and CO2 concentration will have a positive effect on DR and INT (Figure 8(c) and 8(d)). However, the effect of CO2 concentration on DR was not as significant as it is on IRR and ETo. In most places, a unit change in rainfall will result on average in a two-fold increase or decrease in DR in the direction of change in rainfall (Figure 8(c)).
Effect of change in rainfall on the major components of the water budget under different representative concentration pathways (RCPs) compared with the baseline: (a) irrigation requirement (IRR), (b) evapotranspiration (ETo), (c) drainage (DR), and (d) canopy interception (INT).
Effect of change in rainfall on the major components of the water budget under different representative concentration pathways (RCPs) compared with the baseline: (a) irrigation requirement (IRR), (b) evapotranspiration (ETo), (c) drainage (DR), and (d) canopy interception (INT).
CONCLUSIONS
This study investigated the impacts of potential future climate change on citrus water requirements in major citrus producing regions across the world, e.g., Africa (Cape Town, South Africa), Asia (Mersin, Turkey), Australia (Riverland, Australia), Mediterranean (Nabeul, Tunisia), North America (Riverside, California; Fort Pierce and Lake Alfred, Florida; and Brownsville, Texas), and South America (Sao Paulo, Brazil).
Current IRR or water footprint of citrus shows considerable variations across the study sites. While evapotranspiration is dominant in regulating the hydrologic cycle, it is predicted to consistently decrease in all locations, on average, by up to 12 and 11% in the 2055s and 2090s, respectively. Overall, decreases in IRR at greater CO2 concentrations were associated with decreases in ETo, regardless of geographic locations. Under the same CO2 concentration, however, an increase in temperature leads to increases in both ETo and IRR. Canopy interception and drainage will slightly increase under all RCPs at all locations, while runoff will not show a significant change.
While annual IRR is predicted to decrease consistently both during the 2055s and 2090s in all locations, monthly changes in IRR compared with the baseline showed mixed results, especially under RCPs 2.6 and 4.5, where increases in IRR were predicted during some months of the year.
Results from this study underscore the importance of accounting the effects of CO2 concentration in computing evapotranspiration rates. This study also provides insights into how projected increases in atmospheric CO2 concentration and temperature interact and affect major components of the water budget and citrus irrigation requirements. Such results are essential in the global efforts of planning climate change adaptation and mitigation strategies for the citrus industry. However, further studies are needed to investigate how citrus yield would respond under potential climate change, including an economic analysis.
ACKNOWLEDGEMENTS
This work was supported by the USDA National Institute of Food and Agriculture, Evans-Allen project 2014-33100-08916 and project Agrilife-M 1700130 funded by Texas A&M AgriLife.